Background:Biologics for the treatment of rheumatoid arthritis (RA) have different modes of action to target auto-inflammatory processes causing the signs and symptoms of the disease. Different biologics may thus have different effects on inflammatory markers. For instance, previous studies have shown that the interleukin-6-inhibitor tocilizumab (TOC) decreases the level of acute phase reactants (APRs) [1]. Such direct effects on inflammatory markers may lead to an overestimation of clinical response if disease activity is measured via scores including inflammatory markers, such as the Disease Activity Score 28 (DAS 28). The detected changes in disease activity may not adequately reflect the clinical improvement of signs and symptoms.Objectives:In our study, we compared biologics with each other using two different disease activity scores: the DAS 28 including APRs and the clinical disease activity index (CDAI) excluding APRs. The aim of this study was to assess whether the use of the two different scores affects comparative effectiveness studies on biologics for the treatment of RA.Methods:We compared results on the comparative effectiveness of biologics using the corresponding thresholds for low disease activity (LDA) for the DAS 28 (< 3.2) and the CDAI (≤ 10). We performed two separate network meta-analyses (NMAs) after a thorough step-by-step evaluation of the similarity, homogeneity and consistency assumptions of the patient populations and the study data.Our study formed part of a systematic review (including NMAs) that was largely based on clinical study reports and re-analyses of LDA using individual patient data provided by sponsors for studies conducted up to 2017. Thus, the analyses include hitherto unknown data on LDA analysed by means of the CDAI, especially data from older studies. An extensive comparison of DAS 28 and CDAI in different patient populations was possible.Results:For all analysed patient populations, comparisons of TOC versus other biologics yielded remarkable results: advantages for TOC were found in NMAs using the DAS 28, which were not confirmed in NMAs using the CDAI. For methotrexate (MTX)-naïve patients, using the DAS 28, TOC showed a greater benefit than abatacept (ABA), certolizumab pegol (CZP), and etanercept (ETA), which was not confirmed by the CDAI. In contrast, TOC showed less benefit than adalimumab (ADA) and ETA. For patients after MTX failure and using the DAS 28, TOC showed a greater benefit than ABA, ADA, anakinra (ANA), ETA, golimumab (GOL), and infliximab (INF). With the exception of ANA, these advantages were not confirmed by the CDAI. Similar differences between DAS 28 and CDAI were shown in patients treated with biologics in monotherapy or after failure of biologics.Conclusion:In comparative effectiveness studies of biologics, the assessment of LDA using the DAS 28 instead of the CDAI leads to a consistent overestimation of the benefit of TOC in all patient populations, regardless of pre-treatment or combined therapy with MTX. The inclusion of APRs in disease activity scores may thus introduce bias. A score excluding inflammatory markers should therefore be used to ensure valid results.References:[1]Smolen JS, Aletaha D. Interleukin-6 receptor inhibition with tocilizumab and attainment of disease remission in rheumatoid arthritis: the role of acute-phase reactants. Arthritis Rheum 2011; 63(1): 43-52.Disclosure of Interests:Kirsten Janke: None declared, Katharina Biester: None declared, Dietmar Krause Grant/research support from: Pfizer and AbbVie (Abbott), Bernd Richter: None declared, Christoph Schürmann: None declared, Katharina Hirsch: None declared, Beate Wieseler: None declared
The role of matching when adjusting for baseline differences in the outcome variable of comparative effectiveness studies
